KR102473802B1 - Keyword analysis method and online advertisement provision method - Google Patents

Abstract

The present invention relates to a keyword analysis method and, more specifically, to a keyword analysis method and an online advertisement providing method using the same. In accordance with an embodiment of the present invention, the keyword analysis method performed by an electronic device may comprise the following steps of: inputting a keyword related to a product to be sold into a site providing a monthly search number user statistics service, wherein the monthly search number user statistics service includes a statistics service for determining whether a searched keyword is a desktop or mobile device, a gender-based ratio statistics service for the searched keyword, and an age-based ratio statistics service for the searched keyword; in proportion to a ratio between the searched keyword and the desktop and mobile devices, determining marketing costs for the desktop and mobile devices, while, when the searched keyword accounts for 75% or more in the ratio of the mobile device, determining the marketing costs only for the mobile device; receiving keyword information about the product, wherein the keyword information includes a search volume for a specific keyword, the number of competitor products, and a ranking in which the product is exposed with the specific keyword; filtering the keyword information, based on keyword information about at least one role model product; and determining priorities of keywords to be sequentially occupied in regard to the product, based on the filtered keyword information. Therefore, the present invention is capable of establishing an optimal advertising strategy when a keyword is lost to a third party.

Description

Keyword analysis method and online advertisement provision method using it {KEYWORD ANALYSIS METHOD AND ONLINE ADVERTISEMENT PROVISION METHOD}

The present invention relates to a keyword analysis method, and more particularly, to a keyword analysis method and an online advertisement providing method using the same.

With the recent use of the Internet, keyword advertisements among internet advertisements are widely used. Keyword advertising is an advertising technique in which a user inputs a search term (keyword) to expose advertisements of related companies that result in search results. Because feedback can be easily captured through multiple tools, it is a relatively preferred advertising model for marketers.

Recently, the non-face-to-face trend is spreading socially, and according to this trend, the field of online commerce is growing further. For example, in the past 10 years, the net sales of Amazon, a representative online commerce site, have increased by about 9 times from about $ 30 billion to about $ 200 billion. Given that the development of many devices, such as smartphones, is affecting the shopping habits of many people, the field of marketing through the Internet is expected to continue to grow. Keyword advertising is the most important advertising technique in the Internet marketing field.

As such, even if several products of the same product are registered on the Internet, search results displayed on a mobile device or a desktop device to the consumer are inevitably different depending on which keyword the consumer inputs.

Accordingly, if a seller appropriately selects keywords that consumers mainly use when registering a product, the seller will increase the frequency of exposure of the product registered online to customers, thereby increasing sales through online. There is, and market share can be expanded accordingly, and these keywords are purchased through bidding on major main portal platforms/sites (eg, Naver, Kakao, Daum, Google, etc.). For example, when the No. 1 keyword is purchased at A won, A won per click is spent on advertising. It is a structure that becomes

As such, marketing can be determined by purchasing keywords exposed on major main portal platforms/sites through auctions. However, the consumer's purchase process is not simply exposure through keywords, but various psychological factors can determine the exposure and frequency of exposure of keywords in this process, but it is important that the result of such exposure actually leads to consumer purchase. . To this end, it would be very important to diversify marketing strategies according to keywords and consumer statistics.

On the other hand, the above-described background art is technical information that the inventor possessed for derivation of the present invention or acquired during the derivation process of the present invention, and cannot necessarily be referred to as known art to the general public prior to filing the present invention.

Korean Registered Patent No. 10-2127203 Korean Registered Patent No. 10-2194273

The present invention is to solve the above problems, and an object of the present invention is to provide maximum sales within a limited advertising cost, rather than winning a keyword by unconditionally investing excessive costs.

The present invention is not simply bidding for first place in keyword advertisements at an expensive price, but the goal of the present invention is to bid at an appropriate price at a keyword price of a rank displayed on the screen of a mobile device or personal computer (desktop) according to the characteristics of the keyword do it with

Through this, it is possible to enable the seller to obtain an optimal advertising effect without spending a large amount of marketing cost.

The above and other objects and advantages of the present invention will become apparent from the following description of preferred embodiments.

The present invention relates to a keyword analysis method, and more particularly, to a keyword analysis method and an online advertisement providing method using the same. A keyword analysis method, performed by an electronic device according to an embodiment of the present invention, includes the step of inputting a keyword related to a product to be sold to a site providing a monthly search number user statistics service - the monthly search number user statistics service includes: Includes a statistics service for determining whether a searched keyword is a desktop device or a mobile device, a gender-based rate statistics service for the searched keywords, and an age-specific rate statistics service for the searched keywords; The searched keyword determines the marketing cost for desktop devices and mobile devices in proportion to the proportion of desktop devices and mobile devices, and if the keyword searched for is proportional to the proportion of mobile devices, the marketing cost for mobile devices is reduced to the cost for mobile devices. determining only; Receiving keyword information about a product - the keyword information includes a search volume for a specific keyword, the number of competitor's products, and a ranking in which the product is exposed in a specific keyword; Based on keyword information on one or more role model products, filtering keyword information; and determining priorities of keywords to be sequentially occupied with respect to products based on the information of the filtered keywords.

According to an embodiment, when the target target of the keyword is female and the age range is from 20s to 40s, the search gender proportion of the keyword is female in the ratio statistics service according to gender for the searched keyword and the ratio statistics service by age for the searched keyword interpreting that the keyword has a high advertising efficiency score if the ratio exceeds 75% and the upper age group includes people in their 20s to 40s among the groups divided into 10-year-olds; and if the keyword is interpreted as having a high advertisement efficiency score, sending a message inducing the first user to re-bid the keyword.

According to an embodiment, the method may further include displaying a heat map indicating the number of clicks of a specific product page and a exit rate map indicating whether or not a specific Internet page is exited.

According to an embodiment, monitoring whether a keyword for which a first user bid at a first price for a desktop is bid at a second price higher than the first price by a second user - the information about the keyword includes the keyword It includes at least one of group, bid unit, bid price trend, target rank, and competitor's advertising period, and the bid price trend represents a chart displaying the price whenever the owner of the keyword is changed, and the target rank is The ranking displayed on the first screen is set based on the searched keyword, and the competitor's advertising period includes period information about how long users other than the first user preoccupied the keyword before the first user bids on the keyword. Ham -; During monitoring, when a keyword that was bid at a first price by a first user is bid at a second price higher than the first price by a second user, the keyword is submitted to the monthly search user statistics service to determine whether to re-bid the keyword. re-entering; giving up re-bidding at a price higher than the second price used to preoccupy a second user if the search rate of the keyword in the mobile device exceeds 75% at the time when the previously preoccupied keyword is lost; and sending a recommendation message to the first user to run a keyword advertisement through a mobile device instead of a keyword advertisement through a desktop.

According to an embodiment, analyzing purchase propensity by medium, wherein the medium includes banner advertisements, performance-based display advertisements, display networks, and in-stream advertisements; Performing retargeting after analyzing purchase propensity by medium and step-by-step exit rate; Based on the analysis of purchasing propensity for each medium and step-by-step exit rate, the step of determining a retargeting banner advertisement for customers who have departed from the previous month is further included, and the step of prioritizing the keywords determines the advertising efficiency score of the keywords. The step of determining the advertising efficiency score of the keywords includes: inputting a search volume of the keywords; inputting the number of competitor products of keywords; For each of the keywords, calculating an advertising efficiency score by dividing the search volume by the number of competitor's products; and filtering the keyword information, wherein the filtering of the keyword information includes filtering keywords whose search volume of keywords is 10,000 or more and the product is ranked 5th to 10th in a specific keyword.

and the method further comprises separating the filtered keywords at the word level; inputting keywords separated into word levels and monthly search user statistics into a deep learning algorithm for correlation analysis in order to analyze a correlation between keywords separated into word levels and monthly search user statistics; and selecting a keyword having the highest correlation between the keywords derived from the result of the deep learning algorithm and the user statistics of monthly search numbers as a keyword of a new keyword advertisement.

According to the present invention, it is possible to provide an effect that can derive the best advertising effect within a limited advertising budget.

According to the present invention, if there is a concern that the advertising budget will be excessively large and the result of analyzing the corresponding keyword is that the gain is not large, the keyword is given up and another keyword is selected, so that the optimal advertising strategy at the time when the keyword is stolen by a third party enable it to be established.

However, the effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

1 is a diagram schematically illustrating a keyword analysis method according to an embodiment of the present invention.
2 is a diagram illustrating a method of determining an advertisement efficiency score of a current keyword.
3A is an exemplary diagram illustrating a heat map and a bounce rate.
3B is an exemplary diagram showing statistical results.
4 is a diagram illustrating expected effects of growth marketing.
5 is a diagram for explaining a method for retargeting.
FIG. 6 is a site that provides a monthly search number user statistics service and illustrates an exemplary tool for analyzing advertisement performance by major media.
7 is a diagram briefly illustrating the basic concept of an artificial neural network.

Hereinafter, the present invention will be described in detail with reference to embodiments and drawings of the present invention. These examples are only presented by way of example to more specifically describe the present invention, and it will be apparent to those skilled in the art that the scope of the present invention is not limited by these examples.

In addition, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, and in case of conflict, including definitions The description herein will take precedence.

In order to clearly explain the proposed invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification. And, when a certain component is said to "include", this means that it may further include other components without excluding other components unless otherwise stated. Also, a “unit” described in the specification means one unit or block that performs a specific function.

In each step, the identification code (first, second, etc.) is used for convenience of description, and the identification code does not describe the order of each step, and each step does not clearly describe a specific order in context. It may be performed differently from the order specified above. That is, each step may be performed in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

Hereinafter, the configuration of the present invention and its effects will be described in more detail through specific examples and comparative examples. However, these examples are for explaining the present invention in more detail, and the scope of the present invention is not limited to these examples.

In the keyword analysis method according to an embodiment of the present invention, keywords related to products to be sold may be input to a site providing monthly search number user statistics service. Here, the monthly search number user statistics service is a statistics service configured to provide statistical figures as to whether a searched keyword is a desktop device or a mobile device, a ratio statistics service for searched keywords by gender, and a ratio statistics for searched keywords by age group. service may be included.

The searched keyword determines the marketing cost for desktop devices and mobile devices in proportion to the proportion of desktop devices and mobile devices, and if the keyword searched for is proportional to the proportion of mobile devices, the marketing cost for mobile devices is reduced to the cost for mobile devices. can be determined only by Keyword information on products may be received. Recently, customers tend to make more purchases on mobile devices than on desktop devices, mainly because purchases on mobile devices are very convenient and the amount of time spent on mobile devices is high. Here, the keyword information may include a search volume for a specific keyword, the number of competitor's products, and a ranking in which the product is exposed in a specific keyword. When a corresponding keyword is newly occupied by a third party, keyword information may be filtered based on keyword information on one or more role model products. Based on the information of the filtered keywords, it is possible to determine the priority of keywords to be sequentially occupied with respect to the corresponding product. If the auction price of the keywords determined second in priority is 80% or less of the price of the lost keyword, the keyword may be recommended for bidding.

1 is a diagram schematically illustrating a keyword analysis method according to an embodiment of the present invention.

Referring to Figure 1, the analysis method according to an embodiment of the present invention can be performed in the following steps. S110 In step, it may be monitored whether the keyword, which was bid by the first user at the first price, is bid by the second user at a second price higher than the first price. Here, the information on the keyword to be bid may include at least one of a keyword group, a bidding unit, a trend of a bidding price, a target ranking, and an advertisement period of a competitor. A keyword group means a field of a corresponding keyword, and may be, for example, a criterion for determining whether a keyword corresponds to an electronic product group. The bidding unit is usually won, but in the case of overseas platforms (eg, Google, Amazon, etc.), bidding may be conducted in dollars. The trend of the bid price may be a chart displaying the price whenever the owner of the corresponding keyword is changed. The target ranking does not necessarily have to be 1st, and a ranking (eg, 5th) that is displayed on the first screen based on the searched keyword may be set. The competitor's advertisement period may include period information about how long other competitors preoccupied the keyword before a user bids on the corresponding keyword. The above monitoring may be done periodically, and the period may be performed every 2 hours or every 4 hours.

In step S120, when the keyword, which was bid by the first user at the first price, is bid by the second user at a second price higher than the first price, the keyword is used as a monthly search number to determine whether to re-bid the keyword. Can be entered into user statistics service. Here, the monthly search number user statistics service may include a statistics service for determining whether a searched keyword is a desktop device or a mobile device, a statistics service for a searched keyword according to gender, and a statistics service for a searched keyword for each age group. have. Here, the fact that the first user responds only when the keyword is stolen is because it has the effect of reducing the need for continuous management effort.

In a situation where preoccupied keywords are taken away, unconditionally pursuing them by raising the price again can lead to excessive consumption of advertising costs. Therefore, it is essential to check the current status of the advertisement status of the keyword at the time when the preoccupied keyword is taken away. First, in proportion to the ratio of desktop devices and mobile devices of the keyword searched through the automated method according to the invention, and determining a marketing cost for the mobile device. If the ratio of the searched keyword to the mobile device is 75% or more, the marketing cost for the mobile device may be determined only by the cost for the mobile device.

In step S130, if the keyword advertisement was performed only through the desktop in the past, but the search rate on the mobile device exceeds 75% at the time the preoccupied keyword was taken away, the need to re-occupy the keyword related to the desktop that was taken away Since there is no , a recommendation message may be sent to the first user so that the second user gives up re-bidding at a price higher than the second price used to preoccupy, and executes a keyword advertisement through a mobile device rather than a keyword advertisement through a desktop. .

2 is a diagram illustrating a method of determining an advertisement efficiency score of a current keyword.

In step S210, it may be determined whether to change the keyword. In this case, based on the advertisement efficiency score of the current keyword, it may be determined whether to reuse the keyword currently lost to the second user.

In step S220, it is possible to analyze the correlation between the target of the keyword of the corresponding product and information about the statistically appearing keyword. For example, if the target audience of the keyword is female and the age group is in their 20s to 40s, the percentage of searched keywords by gender in the statistics service for searched keywords by gender and the percentage of searched keywords by age group in the statistics service for searched keywords shows that the percentage of searched gender is female. If it exceeds 75% and the top age group includes people in their 20s to 40s, it can be interpreted that the keyword has a high advertising efficiency score (S230), and the first user re-bids the keyword even if he or she spends more advertising money. It is possible to send a message inducing to do (S240). However, if the keyword actually has a high male search rate and does not match the age group, the keyword is interpreted as having a high advertising efficiency score (S250), so there is no need to bid on the current keyword again, and it is difficult to select another keyword. A message recommending a plan may be sent to the first user (S260).

3A is an exemplary diagram illustrating a heat map and a bounce rate map.

In (a) of FIG. 3A , the heat map is a diagram in which the location where the customer clicked is indicated by dots. Through this, it is possible to visually confirm whether the customer actually clicked on a product having a certain keyword.

Through this visual confirmation, it is possible to sort products in descending order to products with the greatest number of hits. Through this, sales of hit products having the greatest number of hits may be further promoted.

In (b) of FIG. 3A , the exit rate map is a gradation display of the degree of deviation of the displayed part. The red color indicates a higher bounce rate, which also indicates that customers do not generally only consider the top-most products, but habitually scroll down.

3B is an exemplary diagram showing statistical results.

Specifically, in FIG. 3B , monthly search trend can be expressed by dividing into desktop and mobile, monthly search user statistics can be displayed as a gender ratio, and monthly search user statistics can be shown by age group.

Figure 3b shows that the core target is set to 30 to 40.

4 is a diagram illustrating expected effects of growth marketing.

According to the present invention, it is possible to maximize efficiency (efficiency optimization, increase in the number of conversions) through an operating strategy using a growth marketing tool for churned customers, non-converting customers, and customers who purchase their own products during advertisement execution.

In digital advertising, KPIs (Key Performance Indicators) can be set for each keyword group. In addition, an efficiency index for each keyword performance can be set.

First, set campaign goals and set goals for each medium.

Set banner media operation strategy. At this time, banner advertisements are executed through media operation through meaningful (customer behavior, action analysis data) of other media.

If there is a customer defection compared to the previous month, a retargeting banner advertisement is performed for the customer who defected. To this end, a process of retargeting may be required after analyzing purchase propensity by medium and step-by-step exit rate analysis.

Advertising methods include banner advertising, performance-based display advertising, display network advertising, and the like.

On the other hand, there is a case where a short advertisement is recently inserted into a video service, and in this case, it is understood that the advertisement efficiency tends to be much higher than that of advertisements through other media. Moreover, since recent video services provide advertisements according to algorithms according to the interests of customers, adding costs for advertisements inserted into video services, that is, in-stream advertisements, can bring practical and efficient effects of advertisements. can be a way for For example, in the case of a YouTube video, a method of charging only when a user watches an advertisement after a 5-second video advertisement played before the video is played is currently being used. That is, in the case of this type of advertisement, it can be considered very efficient if it is suitable for 100% of the searched target and 5 seconds or more of the searched target.

5 is a diagram for explaining a method for retargeting.

In step S510, a product of interest may be exposed on the advertiser's site.

In step S520, the retargeting advertisement may be exposed when the corresponding user visits the media site.

In step S530, the user may be introduced to the advertiser's site to induce purchase.

According to this method, high efficiency can be brought about in retargeting.

In this specification, only a few examples of various embodiments performed by the present inventors are described, but the technical idea of the present invention is not limited or limited thereto, and can be modified and implemented in various ways by those skilled in the art, of course.

FIG. 6 is a site that provides a monthly search number user statistics service and illustrates an exemplary tool for analyzing advertisement performance by major media.

First, Google Analytics can analyze traffic within the website, provide advertisement performance analysis by major media, and monitor incoming media in real time.

In the case of Logger, active IPs can be checked and analyzed, and advertisement performance analysis by major media can be provided.

Ace Counter is a real-time automated bidding solution that can help you establish performance-based bidding strategies.

On the other hand, in order to consider new keywords for advertisements, deep learning algorithms can be used to identify correlations between current keywords, analysis of advertisement performance by major media, and monthly search user statistics. For example, a correlation between specific keywords and a user statistics of monthly searches may be analyzed, and a keyword having the highest correlation may be selected as a keyword of a new keyword advertisement.

A description of the deep learning algorithm applicable to the present invention is as follows.

A deep learning algorithm is one of machine learning algorithms and refers to a modeling technique developed from an artificial neural network by imitating a human neural network. As shown in FIG. 7 , the artificial neural network may be configured in a multi-layered structure.

7 is a diagram briefly illustrating the basic concept of an artificial neural network.

As shown in FIG. 7, an artificial neural network (ANN) is a layer including an input layer, an output layer, and at least one intermediate layer (or hidden layer) between the upper input layer and the output layer. structure can be made. The deep learning algorithm, based on such a multi-layered structure, can derive highly reliable results through learning to optimize the weight of an activation function between layers.

Deep learning algorithms applicable to the present invention may include a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), and the like.

A deep neural network (DNN) is basically characterized by improving the result of learning by increasing a lot of intermediate layers (or hidden layers) in an existing ANN model. As an example, the above DNN is characterized by performing a learning process using two or more intermediate layers.

Accordingly, the computer can derive an optimal output value by repeating the process of generating classification labels, distorting space, and classifying data by itself.

A convolutional neural network (CNN) is characterized by having a structure in which a pattern of features is identified by extracting features of data, unlike existing techniques in which a learning process is performed by extracting knowledge from data. The above CNN may be performed through a convolution process and a pooling process. In other words, the above CNN may include an algorithm complexly composed of a convolution layer and a pooling layer. Here, in the convolution layer, a process of extracting features of data (aka convolution process) is performed. The above convolution process is a process of examining the adjacent components of each component in the data, identifying characteristics, and deriving the identified characteristics into a single sheet. As a single compression process, the number of parameters can be effectively reduced. In the pooling layer, a process of reducing the size of the layer that has undergone the convolution process (aka pooling process) is performed. The above pooling process can reduce the size of data, cancel noise, and provide consistent features in fine parts. For example, the above CNN can be used in various fields such as information extraction, sentence classification, and face recognition.

A recurrent neural network (RNN) is a type of artificial neural network specialized in learning repetitive and sequential data and is characterized by having a recurrent structure therein. The RNN above applies weights to past learning contents using the above circular structure and reflects them to current learning, enabling a connection between current learning and past learning, and has the characteristic of being dependent on time. The above RNN is an algorithm that solves the limitations of existing continuous, iterative, and sequential data learning, and can be used to identify speech waveforms or to identify the front and back components of text.

However, these are only examples of specific deep learning techniques applicable to the present invention, and other deep learning techniques may be applied to the present invention according to embodiments.

On the other hand, it is quite natural for keyword advertisements to determine the bidding price according to the ranking exposed, but in terms of efficiency compared to advertisements, it can be said that there are more cases where this is not the case. Considering this point, the present inventors focused on the fact that even if a preoccupied keyword is stolen by a third party, customers scroll downward at least once rather than spending excessive money to immediately secure the keyword. Thus, a method of bidding for a keyword located at the end that can be exposed through one scroll is presented at a rather cheap price. At this time, the position located at the end that can be exposed through one scrolling may be mouse scrolling in the case of desktop and downward swipe in case of mobile. (e.g., average position in Gaussian distribution), suggesting a way to preoccupy the keyword of the corresponding position, and suggesting a way to obtain the highest efficiency compared to advertising cost.

Claims (5)

In the keyword analysis method, performed by an electronic device,
Entering keywords related to a product to be sold into a site providing monthly search number user statistics service - The monthly search number user statistics service is a statistical service for determining whether a searched keyword is for a desktop device or a mobile device, Includes ratio statistics service by gender for keywords, and ratio statistics service by age for searched keywords -;
determining a marketing cost of a desktop device and a mobile device based on the searched keyword in proportion to the ratio of the desktop device and the mobile device;
When the target target of the keyword is female and the age group is in their 20s to 40s, the gender ratio of the keyword searched in the gender ratio statistics service for the searched keyword and the age ratio statistics service for the searched keyword is more than 75% of the female , interpreting the keyword as having a high advertisement efficiency score if the upper age group includes people in their 20s to 40s among the groups divided by 10 years old; and
If the keyword is interpreted as having a high advertising efficiency score, sending a message inducing the user to re-bid the keyword.
Including, keyword analysis method. delete According to claim 1,
Showing a heat map indicating the number of clicks on a specific product page and a bounce rate map indicating whether or not a specific Internet page is broken
Further comprising a keyword analysis method. According to claim 3,
monitoring whether a keyword, which was bid by a first user at a first price for a desktop, is bid by a second user at a second price higher than the first price - the information about the keyword includes a keyword group, a bidding unit, It includes at least one of a bid price trend, a target rank, and a competitor's advertising period, wherein the bid price trend represents a chart displaying a price whenever the owner of a corresponding keyword is changed, and the target rank represents a searched keyword. It is set as a ranking displayed on the first screen as a standard, and the advertisement period of the competitor includes period information about how long users other than the first user preempted the keyword before the first user bid on the keyword. Ham -;
During the monitoring, when a keyword for which the first user bid at a first price is bid at a second price higher than the first price by a second user, the keyword is used as a monthly search number to determine whether to re-bid the keyword. re-entering the user statistics service;
giving up re-bidding at a price higher than the second price used to preoccupy a second user if the search rate of the keyword in the mobile device exceeds 75% at the time when the previously preoccupied keyword is lost; and
Sending a recommendation message to the first user to run a keyword advertisement through a mobile device instead of a keyword advertisement through a desktop.
Further comprising a keyword analysis method. According to claim 4,
Analyzing purchase propensity by medium - the medium includes banner advertisement, performance display advertisement, display network, and instream advertisement -;
Performing retargeting after analyzing the purchasing propensity for each medium and the step-by-step exit rate;
Determining a retargeting banner advertisement for customers who have lost from the previous month based on the purchasing propensity for each medium and the step-by-step exit rate analysis
Including more,
The step of determining the priority of the keyword,
Determining the advertising efficiency score of the keyword
including,
Determining the advertising efficiency score of the keyword,
inputting a search amount of the keyword;
inputting the number of competitor products of the keyword;
calculating an advertising efficiency score for each of the keywords by dividing the search volume by the number of competitor products; and
Filtering keyword information
including,
Filtering the keyword information,
Filtering keywords whose search volume of the keyword is 10,000 or more and the ranking in which the product is exposed in a specific keyword is ranked 5th to 10th
including,
The method,
Separating the filtered keywords at the word level;
inputting keywords separated at the word level and monthly search user statistics into a deep learning algorithm for correlation analysis in order to analyze correlation between keywords separated at the word level and monthly search user statistics; and
Selecting a keyword having the highest correlation between the keyword derived from the result of the deep learning algorithm and the monthly search user statistics as a keyword of a new keyword advertisement;
Further comprising a keyword analysis method.

Images